search

MCG-NJU / MOTIP

Multiple Object Tracking as ID Prediction
https://arxiv.org/abs/2403.16848
Apache License 2.0
90 stars 8 forks source link

How to run inference on single video? #6

Open tarunmcom opened 5 months ago

tarunmcom commented 5 months ago

How to run inference on single video?

HELLORPG commented 5 months ago

Currently we not support for run inference on a single video (like .mp4 files). This is because the MOT benchmarks store videos in the form of continuous images. If you want to run our model on a video file, a straightforward approach is to sample it into a sequences of images (we suggest about 25 FPS).

In the future, I do plan to implement a demo file like demo.ipynb in MeMOTR, which can directly process a video file. However due to other work arrangements, this planning may require more time. You may be able to modify it based on the demo code of MeMOTR. The code structures of the two are relatively similar.

urbaneman commented 5 months ago

@tarunmcom I added this feature in my own fork project, you can refer to it support video infer

yarkable commented 5 months ago

great! @urbaneman

BianJinLong123 commented 3 months ago

有没有推理的代码呢